In this study, investigation was designed to find an effective method for estimating chlorophyll and nitrogen concentration in the canopies of rice from hyperspectral EO-1 Hyperion image. Continuum-removal analysis enables the isolation of absorption features and minimizes the background influence, thus absorption features stand out. We applied stepwise regression analysis and absorption feature analysis to the field measured foliage and canopy continuum-removed spectra. The results showed that the continuum-removed spectra from the whole range could be broke down into four isolated wavelength ranges and the first wavelength range was centered at 670nm. The area of the wavelength range centered at 670nm based on the BNC spectra was strongly correlated with the chlorophyll and nitrogen concentration. It was validated by EO-1 Hyperion image data, the results showed that the multiple correlation coefficients (R²) between the area of the wavelength range centered at 670nm based on the BNC image spectra and chlorophyll and nitrogen concentration were 0.485 and 0.783 separately. Then the estimation equations were applied to the rice pixels of image which were recognized through Normalized Difference Vegetation Index (NDVI), Land Surface Water Index (LSWI) and Enhanced Vegetation Index (EVI). Thus the chlorophyll and nitrogen concentration distribution maps were obtained. The values in the maps were quite consistent with those of field measurements.

Date Published: 17 October 2006
PDF: 12 pages
Proc. SPIE 6359, Remote Sensing for Agriculture, Ecosystems, and Hydrology VIII, 63591H (17 October 2006); doi: 10.1117/12.689401

Published in SPIE Proceedings Vol. 6359:
Remote Sensing for Agriculture, Ecosystems, and Hydrology VIII
Manfred Owe; Guido D'Urso; Christopher M. U. Neale; Ben T. Gouweleeuw, Editor(s)